[
https://issues.apache.org/jira/browse/TIKA-1085?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Matthew Buckett updated TIKA-1085:
----------------------------------
Attachment: hello-world.pdf
hello-world-bom.pdf
We've seen PDF files with a Byte Order Mark (BOM) prepended to the file. This
has caused Tika to detect the file as text/plain rather than application/pdf. I
can't attach the original files, but attached is an example file
(hello-world.pdf) that is detected correctly and one (hello-world-bom.pdf) that
is incorrectly detected.
All PDF readers are fine opening the file with the BOM. I tried, Chrome's built
in PDF viewer, Preview (Mac) and Acrobat Reader.
> PDF header and mime detection
> -----------------------------
>
> Key: TIKA-1085
> URL: https://issues.apache.org/jira/browse/TIKA-1085
> Project: Tika
> Issue Type: Improvement
> Components: mime
> Affects Versions: 1.3
> Reporter: Marco Quaranta
> Priority: Minor
> Labels: detection, header, mime, pdf
> Attachments: hello-world-bom.pdf, hello-world.pdf, test.pdf
>
>
> I've found some PDF files Tika recognizes as application/octet-stream.
> These files differs from regularly identified PDF having a different header:
> the %PDF-N.n string isn't at the beginning (zero offset) of the file but in
> the first 1024 bytes.
> PDF reference states that "The first line of a PDF file shall be a header
> consisting of the 5 characters %PDF– followed by a version
> number of the form 1.N, where N is a digit between 0 and 7"
> (http://tinyurl.com/8vnzm3c "p. 7.5.2 File Header").
> Looking further at implementation notes by Adobe (http://tinyurl.com/cbqpb24
> p. 3.4.1 File Header) I've discover that: "Acrobat viewers require only that
> the header appear somewhere within the first 1024 bytes of the file"
> What do you think about a PDF magic match with an offset 0:1024?
> <match value="%PDF-" type="string" offset="0:1024"/>
> Thank you,
> Marco
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
